Search

Currently, the huge potential of data centre waste heat for the decarbonization of the heating sector is often ignored. Here, a feasibility study is presented for covering the heat demand of two districts of Frankfurt (Germany) mainly by data centre waste heat. Contrary to many other studies, the total heat demand (144 GWh/a) is to be covered almost exclusively with waste heat integration, the district heating network is not existing yet and the buildings to be connected are not particularly energy efficient. In this study, the potential and demand are estimated, a heating supply concept is presented and evaluated regarding costs, heat pump capacity, and storage size. As a result, the utilisation of data centre waste heat is not only possible, but it is the most promising way for decarbonizing heating in the area under consideration. With high-capacity heat pumps (37 MWth), gas boilers (20 MWth) and a thermal energy storage for daily peaks, 97.5% of the heat demand can be covered with waste heat usage. It is economically favourable compared to individual heat pumps for all types of buildings and with the proposed concept, CO2 emissions in the district heating network area are reduced by 78% on average. [ABSTRACT FROM AUTHOR]

Dynamic thermo-hydraulic simulations of district heating networks (DHN) are essential to investigate novel concepts for their sustainable design and operation. To develop solutions for a particular case study, numerous long-term simulations are required. Therefore, computational effort for simulation is critical. Heat consumers (HC) are numerous and determine the dynamics of mass flows and return temperatures in the DHN. Thus, the way in which HCs are modeled has significant impact on the computational effort and the results of the simulation. This article presents a novel Modelica-based model for HCs that builds on an existing simplified modeling approach (open-loop design). The calculation of mass flow and return temperature is improved in terms of robustness, plausible behavior and low computational effort. In particular, the model reacts to limited differential pressure and supply temperatures to ensure plausible behavior across all operating conditions, including undersupply situations. The model is successfully tested using an exemplary DHN. The analysis proves that the HC model itself requires little time to simulate. Nevertheless, it significantly influences the simulation time for the entire DHN, which varies by a factor of five for the investigated system depending on the HC model. Fast dynamics, including a bypass in the model and correction of deviations between set point and actual heat load increase the simulation time, so users should sensibly choose how to use these options. HC models triggering many state events result in high computational effort. Compared to other simple HC models, the proposed model produces more plausible results while maintaining at least equal simulation performance (for models without bypass) or even improving it (for models with bypass, CPU time is reduced by at least 35%). [ABSTRACT FROM AUTHOR]